# Advantage-Conditioned Supervised Extraction Achieves Optimal RL

- Source: Sergey Levine
- Published: 2026-10-21T04:26:54.000Z
- RoboHot: https://robohot.news/en/items/v75if96pvzyd4kvdaoaxyre83
- Original: https://x.com/svlevine/status/2112762529701405127

## Summary

It turns out that getting a few important details right is all it takes: extracting policies using only advantage-based supervised learning can achieve state-of-the-art results in reinforcement learning (RL)!